【作者】 魏守智； 赵海； 王刚；

【Author】 Shouzhi Wei 1 Hai Zhao 2 Gang Wang2 1(Department of Computer Engineering,Northeastern University at Qinhuangdao,Qinhuangdao,066004) 2(School of Information Science & Engineering,Northeastern University,Shenyang,110004)

【机构】 东北大学秦皇岛分校计算机工程系； 东北大学信息科学与工程学院；

【摘要】 针对丰满水电数字仿真系统的网络在线故障诊断问题,依据信息融合理论,提出了一种新的网络在线智能故障诊断模型。它采用定性与定量、局部与综合诊断相结合的诊断策略,即首先进行快速定性诊断,一旦系统异常,立即启动集成神经网络组对来自系统多侧面的故障特征信息进行定量分析和分类,最后两级D-S证据推理模型在各自的融合中心实现对各子网络的融合,提高了诊断的精度和可靠性。在丰满水电数字仿真系统中设计并实现了智能故障诊断子系统,其成功应用验证了模型的有效性和实际应用价值。更多还原

【Abstract】 In order to solve effectively the issue of network online intelligent faults diagnosis in Fengman hydroelectricity digital simulation system, a kind new intelligent diagnosis model is proposed based on information fusion theories on Internet environment. The integrated project combined qualitative and quantitative diagnosis strategy with local and synthetic strategy. First of all, qualitative diagnosis is performed, in case system is abnormal, the integrated neural networks group will perform immediately quantitative analysis and classification to multi-sources character information from multi-sensors, at last, two level D-S evidence reasoning model will diagnose synthetically to outputs of ANNs and local D-S models in their fusion center respectively, the precision and reliability of diagnosis conclusion are improved evidently. The model has been implemented based on OOP and had been tested and run in Fengman hydroelectricity digital simulation system that we developed, it shows that the proposed model is effective, validated its generality, advantage and practical application value.更多还原